Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
First-Generation Immigrants 1
Original
XXX
Articles
Blackwell
Oxford,
International
IMRE
0197-9183
2007 -
by
UK
Publishing
theMigration
Center
for
LtdReview
Migration
Studies of New York. All rights reserved.
Laura Zimmermann
University of Oxford and IZA
Klaus F. Zimmermann
IZA, Bonn University and DIW Berlin
Amelie Constant
IZA, Georgetown University and DIW DC
INTRODUCTION
As modern societies become more and more multicultural, the concept of
ethnic self-identification, that is, the subjective attachment people have to
ethnic communities, is of increasing importance. International migrants, in
particular, have to reevaluate their connections and allegiance to the host and
origin countries soon after immigration. Empirical research on this topic (see
e.g., Phinney, 1990; Rumbaut, 1994) often considers ethnic self-identification
as a single linear variable, meaning that feelings for the country of origin and
host country are mutually exclusive.
Yet, as has been pointed out in the literature, ethnic self-identification is
a much more complex concept (e.g., Kvernmo and Heyerdahl, 1996; Pirie,
1996; Kinket and Verkuyten, 1997; Kolossov, 1999; Landale and Oropesa,
1Financial
support from Volkswagen Foundation for the IZA project on The Economics and
Persistence of Migrant Ethnicity is gratefully acknowledged. We wish to thank Lilya Gataullina
and one referee for helpful comments on an earlier draft.
2007 by the Center for Migration Studies of New York. All rights reserved.
DOI: 10.1111/j.1747-7379.2007.00093.x
769
770
I M R
2002; Barrington, Herron and Silver, 2003; Bodenhorn and Ruebeck, 2003).
It involves situational feelings of varying degrees. That is, individuals can
alter their feelings over time, place, and surroundings (e.g., Eschbach, Supple,
and Snipp, 1998; Duncan and Trejo, 2005). At the same time, ethnic selfidentification can be partially imposed on individuals externally. Conspicuous
characteristics like ones skin color or hair texture can result in racial labeling
and force people into having feelings that they would not have otherwise
(e.g., Phinney, 1990; Gimnez, 1992; Henry and Bankston, 2001).
This paper acknowledges this complexity and uses an alternative
approach of ethnic self-identification. Instead of a linear concept we propose a
two-dimensional dependent variable that generalizes the various possibilities
immigrants have in forming their ethnic self-identification. Whereas earlier
research mostly focused on some minor aspects of ethnic self-identification
(e.g., Sengstock, 1978; Mouw and Xie, 1999; Farver, Bhadha, and Narang,
2002), we use a more comprehensive range of explanatory variables including
pre- and post-immigration characteristics. As findings for men and women
often differ widely among surveys (e.g., Kinket and Verkuyten, 1997; Khanlou,
2005), we analyze the determinants of ethnic self-identification separately
by gender.
E S-I I
771
TABLE 1
DESCRIPTIVE STATISTICS
Females
Variables
Pre-Immigration
Age at Entry
Religiona
Muslim
Christian
Other religions and nonreligious
No Education in Country of Origina
Country of Origina
Turkey
Ex-Yugoslavia
Southern Europeb
Other countries
Post-Immigration
Years since Immigration
Education in Germany a
No degree
Some degree
Higher degree
Ethnic Identification c
Feel German
Feel connected to the country of origin
Mean
22.7
Males
Standard
Deviation
11.1
Mean
22.2
Standard
Deviation
11.1
0.325
0.612
0.063
0.233
0.469
0.488
0.243
0.423
0.363
0.544
0.094
0.273
0.481
0.499
0.292
0.446
0.342
0.178
0.244
0.236
0.475
0.383
0.430
0.425
0.363
0.170
0.284
0.183
0.481
0.376
0.452
0.387
21.1
10.6
24.0
10.6
0.221
0.276
0.503
0.415
0.396
0.500
0.170
0.328
0.502
0.376
0.489
0.500
2.248
1.465
0.766
0.641
2.184
1.528
0.742
0.652
the men are from southern Europe (Greeks, Italians, and Spaniards). The rest
of the sample contains respondents of other origin countries. While southern
European immigrants are predominantly Christians, only about three quarters
of the Turks are Muslim and 12% are Christians; and about 25% of the
Ex-Yugoslavs are Muslims and about 75% are Christians (not in the table).
(ii) Post-immigration characteristics: The raw statistics in Table 1 show
that the average time living in Germany since immigration is about 21 years
for female and 24 years for male immigrants. While about half of the immigrants have obtained a higher education degree in Germany, 22.1% of the
females and 17.0% of the males do not have any education degree in Germany.
(iii) Ethnic identification variables: While our thesis is that ethnic selfidentification should be analyzed on a more complex basis than just a linear
I M R
772
TABLE 2
CROSS-TABULATION OF THE ATTACHMENT TO GERMANY AND THE COUNTRY OF ORIGIN
Feel Connected to the Country of Origin
Females
Feel German
3
2
1
Males
227
37.46%
103
17.00%
43
7.10%
31
5.12%
104
17.16%
49
8.09%
13
2.15%
7
1.16%
29
4.79%
197
30.78%
119
18.59%
42
6.56%
40
6.25%
132
20.63%
54
8.44%
9
1.41%
15
2.34%
32
5.00%
Notes: Feel German: To what extent do you view yourself as a German? = 1, if completely or for the most part;
= 2, if in some respects; = 3, if hardly at all or not at all.
Feel connected to the country of origin: To what extent do you feel that you belong to the culture of the country
where you or your family comes from? = 1, if to a very large extent or to a large extent; = 2, if in some
respects; = 3, if hardly or not at all.
Number of observations: 606 females and 640 males.
linear model assumes a rather restrictive framework, where self-identification with the host
and origin country are mutually exclusive.
E S-I I
773
equals 1) but having only a weak connection to the country of origin (feel
connected to the origin equals 3). As these tabulations show, 59.41% of
the females and 56.41% of the males in our sample exhibit an ethnic selfidentification that is consistent with the concept of mutual exclusiveness. The
rest of the cells, however, show that a considerable percentage of immigrants
experience a more complex perception of their identity. Specifically, they selfidentify with both countries and cultures, and implicitly use a two-dimensional
approach to the ethnic self-identification. This observation warrants our study
and promises interesting results. What determines which path immigrants
choose in self-identifying after arriving in the host country? There are three
possible paths of adjustment from separation at entry, namely, the transitions
to assimilation, integration, and marginalization.
Similar to theories and models on ethnic identity (Berry, 1980), we assume
that there are four two-dimensional possibilities of how people view their ethnic
self-identification, as they embrace new or shed old ethnic identities: (i) Immigrants can be classified as integrated, that is, they feel strongly connected to
both the host and origin country, and these feelings can peacefully coexist;
(ii) Immigrants can be assimilated, meaning that they completely adapt to
and identify with the host country and disengage from the country of origin;
(iii) Immigrants can be labeled as separated, in the sense that they maintain
strong connections to the country of origin, and exhibit or develop only a weak
link with the host country; or (iv) immigrants can go to an extreme situation
where they have loose connections to both the host and the origin country, and
then they are marginalized.
In contrast to these approaches, however, here we are mainly interested
in what makes people deviate from the linear model. Our dependent variable
for ethnic self-identification incorporates the different possibilities of bonding
and attachment to the host and origin countries. This variable has three categories. Category 1 includes those respondents who, right after immigration,
are clearly on their way from separation to marginalization. These are those
immigrants who not only do not adjust to the German way of living, style, and
ethnicity but are also disconnecting with their origin country ethnicity (see
cells (3,2) and (3,3) in Table 2). We also classify people as marginalized if they
are hardly at all or no longer affiliated with the country of origin, but in some
respects they feel German (see cell (2,3) in Table 2).
In category 2 we collect those respondents who are on their direct
way from separation to assimilation, and therefore exhibit their ethnic selfidentification as a linear model. These individuals cluster along the main
diagonal of Table 2, namely in cells (3,1), (2,2), and (1,3). They all behave
774
I M R
consistently with the linear model by identifying themselves either with the
ethnicity of the origin country (cell (3,1)), which is the state of separation, with
the ethnicity of the host country (cell (1,3)), which is the state of assimilation,
or they are on their way from separation to assimilation as attached to both
ethnicities in some respects (cell (2,2)).
Category 3 contains those respondents who are on their way from
separation to integration. It includes immigrants who strongly identify with
one ethnicity and moderately with the other as in cells (2,1) and (1,2) and
those who identify strongly with both ethnicities, namely, individuals from cell
(1,1), who are fully integrated.
To sufficiently cover the determinants of ethnic self-identification, we
employ important pre- and post-immigration characteristics. The preimmigration characteristics include the age at entry into Germany, dummy
variables for Muslim and Christian religion (with individuals of other religions
and nonreligious people being the reference group), no education in the origin
country (the reference group is some kind of education in the origin country),
ex-Yugoslavia, southern Europe (Greece, Italy, and Spain), and people from other
origin countries (with the reference group being Turkey). The post-immigration
characteristics include the continuous variable years since immigration, and
the dummy variables no education degree in Germany and higher degree in
Germany (with some education degree in Germany as the reference category).
In analyzing ethnic self-identification, we expect to find gender
differences in line with Kinket and Verkuyten (1997), and especially
Khanlou (2005), who provide a good overview of research results on ethnic
self-identification and gender differences. While there are gender differences in
most of the studies, the direction of the differences varies. How female and
male immigrants differ in our model is therefore worth pursuing.
EMPIRICAL RESULTS
In order to analyze the determinants of ethnic self-identification in category 1
(marginalized), category 2 (separated, undefined, assimilated), and category 3
(thoroughly integrated), we first employ the multinomial logit model, which
allows for an unstructured and flexible specification.3 In this exercise we
disaggregate again by gender. We use likelihood-ratio tests to measure the value
of pre- and post-immigration characteristics against the full model and against
the base model with a constant only.
3 The
E S-I I
775
TABLE 3
LIKELIHOOD-RATIO TESTS OF THE EFFECT OF PRE- AND POST-MIGRATION CHARACTERISTICS
Likelihood-Ratio Test against 0
Effect of pre-immigration characteristics
Effect of post-immigration characteristics
Effect of full model
Likelihood-Ratio Test against the Full Model
Effect of pre-immigration characteristics
Effect of post-immigration characteristics
Females
Males
49.407***
(7)
17.80***
(3)
61.40***
(10)
50.83***
(7)
23.93***
(3)
63.03***
(10)
43.61***
(7)
12.01***
(3)
39.12***
(7)
12.21***
(3)
Notes: Table shows chi-squared values with degrees of freedom in parentheses; bases of these results are the corresponding
multinomial logit models.
***significant at 99.5%.
TABLE 4
THE DETERMINANTS OF ETHNIC SELF-IDENTIFICATION (ORDERED AND BINARY PROBIT MODELS)
Females
Pre-Immigration
Age at Entry
Muslim
Christian
Southern Europe
Other Countries of Origin
Post-Immigration
Years since Immigration
No Degree in Germany
Higher Degree in Germany
Constant (probit)
1 (ordered probit)
2 (ordered probit)
Log- likelihood
Mean dep. var. (SD)
Pseudo-R2
Marginalization
(Binary Probit)
0.010*
(1.66)
0.472*
(1.96)
0.156
(0.67)
0.191
(1.12)
0.012
(0.07)
0.114
(0.66)
0.261
(1.54)
0.007
(0.87)
0.082
(0.27)
0.020
(0.07)
0.028
(0.11)
0.171
(0.78)
0.725***
(2.71)
0.367
(1.56)
0.009*
(1.69)
0.255
(1.23)
0.121
(0.59)
0.142
(0.95)
0.052
(0.37)
0.303*
(1.95)
0.296*
(1.99)
0.012*
(2.03)
0.311*
(1.65)
0.303*
(2.01)
0.820**
(2.52)
0.001
(0.15)
0.355
(1.39)
0.452*
(2.16)
0.994*
(2.27)
356.1
0.322
(0.468)
0.065
166.5
0.084
(0.278)
0.049
0.008
(1.44)
0.338*
(2.07)
0.358***
(2.69)
1.060
0.876
511.6
2.238
(0.592)
0.043
Integration
(Binary Probit)
Marginalization
(Binary Probit)
Ethnic
Self-Identification
(Ordered Probit)
0.003
(0.50)
0.478**
(2.46)
0.055
(0.29)
0.053
(0.33)
0.206
(1.34)
0.032
(0.20)
0.296*
(1.82)
0.001
(0.16)
0.087
(0.36)
0.133
(0.55)
0.171
(0.76)
0.238
(1.15)
0.513*
(2.10)
0.228
(1.14)
0.003
(0.52)
0.259
(1.52)
0.023
(0.14)
0.082
(0.57)
0.206
(1.53)
0.139
(0.97)
0.085
(0.59)
0.015***
(2.69)
0.181
(1.09)
0.157
(1.16)
0.541*
(1.84)
0.012
(1.63)
0.230
(0.98)
0.067
(0.38)
0.740*
(1.92)
0.014***
(2.95)
0.047
(0.33)
0.083
(0.69)
391.1
0.336
(0.473)
0.043
193.8
0.100
(0.300)
0.068
1.062
0.706
571.5
2.236
(0.617)
0.029
I M R
No Education in
Country of Origin
Ex-Yugoslavia
Integration
(Binary Probit)
776
Variables
Males
Ethnic
Self-Identification
(Ordered Probit)
E S-I I
777
exercise. We find that for women age at entry has a negative effect on ethnic
self-identification. The older immigrant women are when entering Germany,
the less likely they are to self-identify as integrated, and the more likely they are
to identify as members of the linear approach or as marginalized. Religion
and education in the country of origin have no significant effect. Ethnicity or
country of origin, on the other hand, has an effect on the southern immigrants
self-identification and on that of those from other origin countries: compared
to Turks, they both are more likely to self-identify as integrated.
Education in Germany is also of grave importance. Interestingly, both
having no degree and having a higher degree in Germany make immigrants
feel more attached to Germany and more likely to self-identify as integrated,
compared to having some degree. While years since immigration is the only
significant variable for males, it is in line with theory and intuition. The coefficients show that the longer immigrant men live in Germany, the more
likely they are to self-identify as integrated, and the less likely they are to
identify as marginalized.
As the characteristics of the people who differ from the linear approach
are especially interesting, we also employ binary probit models with a minor
change on the dependent variable. Here, the dependent variable integration
has two categories: category 1 contains all those people who were identified as
belonging to category 3 of the variable ethnic self-identification, and category
0 contains all other respondents. This dependent variable therefore captures
the characteristics of the people who feel integrated in comparison to the rest
of the sample. The dependent variable marginalization also has two categories:
category 1 contains all those people who identified themselves as belonging
to category 1 of the variable ethnic self-identification and category 0 contains
all other respondents. This dependent variable therefore analyzes the characteristics of the people who feel marginalized in comparison to the rest of
the sample.
Table 4 presents the results of the binary probit models for the dependent
variables integration and marginalization for both women and men.5 We first
report the findings on the female sample. Clearly, the older women are when
the horizons and results in positive and productive individuals. In terms of the Venn diagram,
for example, integration covers the entire area of both circles. In contrast, assimilation indicates
that immigrants conform to the norms, assume a new personality, that of natives, and have
nothing new to offer to the new society. Marginalization of course indicates disgruntled and
confused individuals, with potentially unintended social ramifications.
5
The descriptive statistics for both endogenous variables are contained in Table 4.
778
I M R
entering the host country, the less likely they are to self-identify as integrated.
Having no degree or a higher degree in Germany both determine and lead to
more integration, compared to having some degree in Germany. In addition,
Muslims are less likely to reveal themselves as integrated than those of other
religions and nonreligious people, whereas there is no such difference for
Christians. The more time has passed since immigration, the more likely
immigrants are to identify as integrated.
Regarding the marginalization category, our results show that southern
European immigrants and those with a higher degree in Germany are less likely
to be found in the marginalized group than Turks or those with some degree
in Germany; all other countries of origin, religion, and education in the country
of origin play no significant role. This means that important pre-immigration
characteristics play no role in determining how people self-identify in the host
country.
For the male sample, Table 4 further shows that being Muslim makes
immigrants less likely to self-identify as integrated than people of other
religions or nonreligious immigrants. People of other ethnicities are more
likely than Turks to self-identify as integrated. Also, Table 4 clearly demonstrates the positive effect of residence in Germany on feelings of integration:
the longer immigrant men have been living in Germany, the more likely they
are to feel integrated. In the case of marginalization, the only variable that has
a significant impact is southern European origin: people from southern Europe
are less likely to feel marginalized in Germany than Turks.
E S-I I
779
780
I M R
REFERENCES
Barrington, L. W., E. T. Herron, and B. D. Silver
2003 The Motherland Is Calling: Views of Homeland among Russians in the Near Abroad.
World Politics 55(2):290 313.
Berry, J.
1980 Acculturation as Varieties of Adaptation. In Acculturation: Theory, Models and Some New
Findings. Ed. A. M. Padilla. Boulder CO: Westview. Pp. 925.
Bodenhorn, H., and C. S. Ruebeck
2003 The Economics of Identity and the Endogeneity of Race. NBER Working Paper no. 9962.
Boston: National Bureau of Economic Research.
Duncan, B., and S. Trejo
2005 Ethnic Identification, Intermarriage, and Unmeasured Progress by Mexican Americans.
IZA Discussion Paper no. 1629. Bonn: Institute for the Study of Labor.
Eschbach, K., K. Supple, and M. Snipp
1998 Changes in Racial Identification and the Educational Attainment of American Indians,
19701990. Demography 35(1):35 43.
Farver, J. A., B. R. Bhadha, and S. K. Narang
2002 Acculturation and Psychological Functioning in Asian Indian Adolescents. Social
Development 11(1):1129.
Gimnez, M. E.
1992 U.S. Ethnic Politics: Implications for Latin Americans. Latin American Perspectives
19(4):717.
Henry, J. M., and Bankston III, C. L.
2001 Ethnic Self-Identification and Symbolic Stereotyping: The Portrayal of Louisiana
Cajuns. Ethnic and Racial Studies 24(6):10201045.
Khanlou, N.
2005 Cultural Identity as Part of Youths Self-Concept in Multicultural Settings. International
Journal of Mental Health and Addiction 3(2):114.
Kinket, B., and M. Verkuyten
1997 Levels of Ethnic Self-Identification and Social Context. Social Psychology Quarterly
60(4):338354.
Kolossov, V.
1999 Ethnic and Political Identities and Territorialities in the Post-Soviet Space. GeoJournal
48:7181.
Kvernmo S., and S. Heyerdahl
1996 Ethnic Identity in Aboriginal Sami Adolescents: The Impact of the Family and the
Ethnic Community Context. Journal of Adolescence 19(5):453463.
Landale, N. S., and R. S. Oropesa
2002 White, Black, or Puerto Rican? Racial Self-identification among Mainland and Island
Puerto Ricans. Social Forces 81(1):231254.
Mouw, T., and Y. Xie
1999 Bilingualism and the Academic Achievement of First- and Second-Generation Asian
Americans: Accommodation with or without Assimilation? American Sociological Review
64:232252.
E S-I I
781
Phinney, J. S.
1990 Ethnic Identity in Adolescents and Adults: Review of Research. Psychological Bulletin
180(3):499514.
Pirie, P. S.
1996 National Identity and Politics in Southern and Eastern Ukraine. Europe-Asia Studies
48(7):10791104.
Rumbaut, R. G.
1994 The Crucible Within: Ethnic Identity, Self-Esteem, and Segmented Assimilation among
Children of Immigrants. International Migration Review 28(4):748794.
Sengstock, M. C.
1978 Developing an Index of Ethnic Community Participation. International Migration
Review 12(1):55 65.
SOEP Group
2001 The German Socio-Economic Panel (GSOEP) After More Than 15 Years Overview.
Ed. E. Holst, D. R. Lillard, and T. A. DiPrete. In Proceedings of the 2000 Fourth International Conference of German Socio-Economic Panel Study Users (GSOEP 2000),
Vierteljahreshefte zur Wirtschaftsforschung (Quarterly Journal of Economic Research)
70:714.